Protracted viral shedding and viral load are associated with ICU mortality in Covid-19 patients with acute respiratory failure

Abstract

Background: Protracted viral shedding is common in hospitalized patients with COVID-19 pneumonia, and up to 40% display signs of pulmonary fibrosis on computed tomography (CT) after hospital discharge. We hypothesized that COVID-19 patients with acute respiratory failure (ARF) who die in intensive care units (ICU) have a lower viral clearance in the respiratory tract than ICU patients discharged alive, and that protracted viral shedding in respiratory samples is associated with patterns of fibroproliferation on lung CT. We, therefore, conducted a retrospective observational study, in 2 ICU of Lyon university hospital. Results: 129 patients were included in the study, of whom 44 (34%) died in ICU. 432 RT-PCR for SARS-CoV-2 were performed and 137 CT scans were analyzed. Viral load was significantly higher in patients deceased as compared to patients alive at ICU discharge (p < 0.001), after adjustment for the site of viral sampling and RT-PCR technique. The median time to SARS-CoV-2 negativation on RT-PCR was 19 days [CI95 %:15–21] in patients alive at ICU discharge and 26 days [CI95 %:17-infinity] in non-survivors at ICU discharge. Competitive risk regression identified patients who died in ICU and age as independent risk factors for longer time to SARS-CoV-2 negativation on RT-PCR, while antiviral treatment was independently associated with shorter time. None of the CT scores exploring fibroproliferation (i.e., bronchiectasis and reticulation scores) were significantly associated with time to SARS-CoV-2 negativation. Conclusions: Viral load in respiratory samples is significantly lower and viral shedding significantly shorter in ICU survivors of COVID-19 associated acute respiratory failure. Protracted viral shedding is unrelated to occurrence of fibrosis on lung CT.